[R] [R} Getting 'tilting' confidence intervals in R

Robert A. LaBudde ral at lcfltd.com
Thu Oct 18 19:11:50 CEST 2007


[Resend with proper subject line]

I am trying to compute bootstrap confidence intervals for a sample 
using R 2.5.1 for Windows.

I can get "Normal", "Basic", "Percentile", "BCa" and "ABC" from 
boot.ci() and boot() in the Davison & Hinkley "boot" package.

But I can't figure out how to use tilt.boot() to get the "tilting" 
confidence interval. Here's a program snippet:

g = rgamma(N,shape=2,scale=3) #Generate a sample of N random deviates

varf = function (x,i) { var(x[i]) }
b = boot(g, varf, R=1000)
boot.ci(b)

fabc = function (x, w) { sum(x^2*w)/sum(w)-(sum(x*w)/sum(w))^2 } #wgt 
average (biased) variance
abc.ci(g, fabc) #ABC method confidence interval

bt = tilt.boot(g, varf, R=c(1000,1000,1000))

The bt object doesn't have a confidence interval method or property, 
even though alpha=c(.025, .975) is a default parameter in the tilt.boot() call.

This must be simple, but I'm not finding out the answer from the documentation.

Also, use of any other packages to accomplish the tilting interval 
would also be useful.

Thanks.



More information about the R-help mailing list